0
TECHNICAL PAPERS

High Performance Motion Tracking Control for Electronic Manufacturing

[+] Author and Article Information
Benjamin Potsaid

Center for Automation Technologies and Systems (CATS), Rensselaer Polytechnic Institute, Troy, NY 12180potsab@rpi.edu

John Ting-Yung Wen1

Center for Automation Technologies and Systems (CATS), Rensselaer Polytechnic Institute, Troy, NY 12180wenj@rpi.edu

Mark Unrath

 Electro Scientific Industries, Inc., Portland, OR 97229unrathm@esi.com

David Watt2

 Electro Scientific Industries, Inc., Portland, OR 97229david.watt@sri.com

Mehmet Alpay

 Electro Scientific Industries, Inc., Portland, OR 97229alpaym@esi.com

1

Also at Department of Electrical, Computer, Systems Engineering.

2

Also at SRI International.

J. Dyn. Sys., Meas., Control 129(6), 767-776 (Mar 02, 2007) (10 pages) doi:10.1115/1.2789467 History: Received April 06, 2006; Revised March 02, 2007

Motion control requirements in electronic manufacturing demand both higher speeds and greater precision to accommodate continuously shrinking part/feature sizes and higher densities. However, improving both performance criteria simultaneously is difficult because of resonances that are inherent to the underlying positioning systems. This paper presents an experimental study of a feedforward controller that was designed for a point-to-point motion control system on a modern and state of the art laser processing system for electronics manufacturing. We systematically apply model identification, inverse dynamics control, iterative refinement (to address modeling inaccuracies), and adaptive least mean square to achieve high speed trajectory tracking. The key innovations lie in using the identified model to generate the gradient descent used in the iterative learning control, encoding the result from the learning control in a finite impulse response filter and adapting the finite impulse response coefficients during operation using the least-mean-square update based on position, velocity, and acceleration feedforward signals. Experimental results are provided to show the efficacy of the proposed approach, a variation of which has been implemented on the production machine.

Copyright © 2007 by American Society of Mechanical Engineers
Your Session has timed out. Please sign back in to continue.

References

Figures

Grahic Jump Location
Figure 1

Feedback and feedforward control architecture

Grahic Jump Location
Figure 2

Laser scanner configuration and experimental testbed. The f-θ lens provides a flat image field and a linear relationship between x, y coordinates and θx, θy, respectively.

Grahic Jump Location
Figure 3

Pole/zero comparison between zero-delay and 16-sample delay identified models

Grahic Jump Location
Figure 4

Gain/phase comparison between experimental data and identified model with 16-sample delay

Grahic Jump Location
Figure 5

Step response comparison between experimental data and identified model with 16-sample delay

Grahic Jump Location
Figure 6

Inverse dynamics feedforward control architecture using the identified model

Grahic Jump Location
Figure 7

Experimental results using the inverse dynamics filter for a 500μm move length

Grahic Jump Location
Figure 8

Inverse dynamics combined with iterative refinement feedforward control architecture

Grahic Jump Location
Figure 9

Procedure of calculating G*e

Grahic Jump Location
Figure 10

Experimental results of applying iterative refinement to 500μm and 1000μm moves

Grahic Jump Location
Figure 11

Feedforward control architecture using a combination of inverse dynamics and the FIR filter

Grahic Jump Location
Figure 12

Command input comparison between inverse dynamics, iterative refinement, and FIR filter: 500μm move

Grahic Jump Location
Figure 13

Output comparison between iterative refinement and FIR filter: 500μm move

Grahic Jump Location
Figure 14

Feedforward control architecture using a combination of inverse dynamics and adaptive FIR filtering

Grahic Jump Location
Figure 15

Comparison between fixed coefficient FIR, LMS updating position only, LMS updating position and velocity, and LMS updating position, velocity, and acceleration

Grahic Jump Location
Figure 16

Desired output trajectory

Grahic Jump Location
Figure 17

Tracking error for random moves: inverse dynamics versus FIR filter

Grahic Jump Location
Figure 18

Tracking error for random moves: comparison between fixed coefficient FIR, LMS updating position only, LMS updating position and velocity, and LMS updating position, velocity, and acceleration

Tables

Errata

Discussions

Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging and repositioning the boxes below.

Related Journal Articles
Related eBook Content
Topic Collections

Sorry! You do not have access to this content. For assistance or to subscribe, please contact us:

  • TELEPHONE: 1-800-843-2763 (Toll-free in the USA)
  • EMAIL: asmedigitalcollection@asme.org
Sign In